Method For Determining Fabric And Upscaled Properties Of Geological Sample

1. A method for determining fabric of a geological sample, comprising: a) detecting key points for at least one digital image of a geological sample; b) feature extracting using the key points to quantitatively represent resolved and unresolved texture using a model; c) pattern recognizing for feature classification which defines clusters of key points corresponding to a perceptually homogenous texture; d) texture segmenting to construct resolved and unresolved area/volume based on the clusters of key points.

2. The method of claim 1 , wherein the feature extracting in step b) comprises: i) discrete wavelet decomposing of the at least one digital image using a discrete wavelet transform to create images at different resolution; ii) computing pixel/voxel gradient of each image obtained from step i) to create gradient images; iii) computing Laplacian-of-Gaussian (LoG) on the gradient images to locate points of variation of pixel/voxel intensity; iv) key point detecting within rock fabrics by locating maxima in the gradient images on which LoG is computed in iii); v) eliminating key points at edges of features of the geological sample in the gradient images on which LoG is computed in iii).

3. The method of claim 2 , wherein the eliminating of key points at edges of the geological sample comprises eliminating key points within 80% of the maxima.

4. The method of claim 2 , wherein the eliminating of undesired key points in step v) comprises eliminating key points at edges and corners of the gradient images.

5. The method of claim 1 , wherein the pattern recognizing uses machine learning to develop a mathematical model for fabric classes (clusters).

6. The method of claim 5 , wherein the texture segmenting comprises extracting features from every pixel and comparing the extracted features to the mathematical model for fabric classes wherein each of the pixels is assigned to one of the fabric classes.

7. The method of claim 1 , wherein the at least one digital image having key points detected in step a) is at least one three-dimensional image of the geological sample.

8. The method of claim 1 , wherein the geological sample is a rock sample.

9. The method of claim 8 , wherein the rock sample is selected from micro-cores, core pieces, drill cuttings, sidewall cores, outcrop quarrying, or whole intact rocks.

10. A system for determining fabric of a geological sample, comprising: i) a digital image acquisition system for obtaining digital images on a geological sample, wherein the digital image acquisition system is selected from at least one of X-ray computed tomography (CT) scanner, micro X-ray CT, nuclear magnetic resonance (NMR) imaging system, scanning electron microscope (SEM), and focused ion beam-scanning electron microscope (FIB-SEM); ii) one or more computer systems comprising at least one processing unit and/or computer programs stored on a non-transitory computer-readable medium operable to a) capture digital images obtained by the digital image acquisition system, b) detect key points for at least one digital image of a geological sample, c) extract features using the key points to quantitatively represent resolved and unresolved texture using a model, d) recognize a pattern for feature classification which defines clusters of key points corresponding to a perceptually homogenous texture, and e) segment texture to construct resolved and unresolved area/volume based on the clusters of key points.